Journal of Family Business Strategy 5 (2014) 347357

Contents lists available at ScienceDirect

Journal of Family Business Strategy

journal homepage: www.elsevier.com/locate/jfbs

Tax aggressiveness in private family rms: An agency perspective

Tensie Steijvers a,*, Mervi Niskanen b
a
b

KIZOK Research Center, Hasselt University, Agoralaan Building D, 3590 Diepenbeek, Belgium
University of Eastern Finland, School of Business, Kuopio, Finland

A R T I C L E I N F O

A B S T R A C T

Article history:
Received 8 March 2013
Received in revised form 16 April 2014
Accepted 9 June 2014

This article investigates, from an agency perspective, whether private family rms, compared to private
nonfamily rms, are more tax aggressive. Moreover, for private family rms, the effect of the extent of
separation between ownership and management on tax aggressiveness is studied. Additionally, we
verify whether effective board monitoring moderates this relationship. Using Finnish survey data, results
show that private family rms are less tax aggressive than nonfamily rms. For the subsample of private
family rms, rms with a lower CEO ownership share are more tax aggressive whereas the presence of an
outside director in their board mitigates this direct effect.
2014 Elsevier Ltd. All rights reserved.

Keywords:
Tax aggressiveness
Family rms
CEO ownership
Board of directors

Introduction
Accounting practices in (private) family rms are rarely studied
(Salvato & Moores, 2010), even though accounting research is one
of the eldest business disciplines and family business represents
the prevalent form of economic organization in the world
(Songini, Gnan, & Malmi, 2013, p. 71). With respect to the
accounting topic of our study, tax aggressiveness, current scant
literature mainly focuses on public family rms and how their use
of tax aggressiveness differs from public nonfamily rms (for
example Chen, Chen, Cheng, & Shevlin, 2010). Whether tax
aggressiveness also prevails within private family rms and
how this tax aggressive behaviour can differ within the heterogeneous group of private family rms remains unstudied.
However, private family rms are characterized by an
entanglement of the family throughout the organization which
affects the nature and extent of agency conicts within the family
rm and is expected to affect the managements tax aggressive
behaviour. Tax aggressiveness is dened as downward management of taxable income through tax planning activities which can
be legal or illegal or may lie in between (Frank, Lynch, & Rego,
2009). Recent evidence shows that management engaging in tax
aggressive activities to minimize tax payment is becoming an
increasingly common feature of the corporate landscape around
the world (Lanis & Richardson, 2011). Desai and Dharmapala

* Corresponding author. Tel.: +32 11 268627; fax: +32 11 268700.

E-mail addresses: tensie.steijvers@uhasselt.be (T. Steijvers),
mervi.niskanen@uef. (M. Niskanen).
http://dx.doi.org/10.1016/j.jfbs.2014.06.001
1877-8585/ 2014 Elsevier Ltd. All rights reserved.

(2006) indicate that the analysis of a tax aggressiveness decision is

embedded in an agency framework in which managers can enjoy
private benets of control at the expense of other shareholders. As
the CEO plays an economically signicant role in determining the
level of tax avoidance that rms undertake, the CEO is the key
driver of corporate behaviour (Dyreng, Hanlon, & Maydew, 2010;
Hambrick & Mason, 1984; Zona, Minoja, & Coda, 2013). To
determine the level of tax aggressiveness a family rm decides to
engage in, the CEO will trade off the marginal benets against the
marginal costs of managing taxes (Molero & Pujol, 2012).
For private family rms, the benets do not only include the tax
savings. Critical characteristics of tax aggressive activities are
complexity and obfuscation. Such a complexity can allow the CEO
to mask any kind of rent extraction vis-a`-vis the other shareholders
(for example perquisite consumption and excessive salaries). This
rent extraction can be considered as agency costs for the rm. On
the cost side, the CEO has to take into account the time that has to
be invested to implement the tax evasion measures, not only the
possible penalty from tax authorities harming his own reputation,
but also the possible damage to the rms reputation and familys
socioemotional wealth (SEW) which is a key noneconomic
reference point for decision making (Berrone, Cruz, Gomez-Mejia,
& Larraza-Kintana, 2010). SEW represents noneconomic goals
(Chrisman, Chua, Pearson, & Barnett, 2010) such as preservation of
the family dynasty and perpetuation of family values through the
business that meet the familys affective needs (Gomez-Mejia,
Haynes, Nunez-Nickel, Jacobson, & Moyano-Fuentes, 2007). Private
family rms have a much longer investment horizon and greater
reputation concerns (Gedajlovic & Carney, 2010) indicating that
they do not only have nancial goals. If the family rm engages in

348

T. Steijvers, M. Niskanen / Journal of Family Business Strategy 5 (2014) 347357

tax aggressive behaviour, reputational damage cannot only occur

due to tax-penalties, reported upon in the press, but also due to the
aggressive usage of legal measures that corporations take to avoid
taxes. Tax returns of all companies and individuals are public
information in Finland, the context of our study. When this
information is released by the tax authorities, the press can publish
the information as well as investigate any deviations from the
norm. However, given the large amount of private Finnish rms,
this will not occur systematically.
Chen et al. (2010) indicate that our understanding of the
determinants of tax reporting aggressiveness is limited. This
literature is relatively young and therefore, most studies have only
examined rm specic determinants using a number of proxies
such as rm size, leverage, scale of operations. . . without much
examination of executives and their incentives (Hanlon & Heitzman, 2010). Shackelford and Shevlin (2001), Scholes, Wolfson,
Erickson, Maydew, and Shevlin (2005) and Desai and Dharmapala
(2006) call for more research of tax management in the presence of
agency conicts. Based on Chen et al. (2010) who study the impact
of public family ownership on tax aggressiveness, we investigate
from an agency perspective whether private family rms,
compared to private nonfamily rms, are more or less eager to
engage in tax aggressive behaviour. Moreover, we will study
whether the extent of separation between ownership and
management, affecting the extent of agency problems, will also
affect tax aggressive behaviour. Additionally, we extend prior
knowledge by studying how effective monitoring by a board of
directors may mitigate the agency problems arising from separation between ownership and control, resulting in tax aggressive
behaviour.
In our study, we use Finnish data as Finland belongs to the
group of high tax alignment countries like for example France and
Spain (Van Tendeloo & Vanstraelen, 2008). High tax alignment
means that there is a high alignment between nancial reporting
and tax accounting. While the general rule is that all the revenues
and expenses have to be reported identically in the tax returns and
the ofcial nancial statements, there are some exceptions. These
can be applied in family as well as nonfamily rms. According to
Atwood, Drake, Myers, and Myers (2012) tax aggressive strategies
can be dened as those that create permanent or temporary booktax differences as well as those that create no differences. As for
permanent tax avoiding strategies on the revenue side, the most
important exceptions are that revenues received from the sale of
shares listed in the rms permanent assets and dividends received
from other companies are tax exempt. This has led to a situation,
where setting up group structures has become a popular tax
planning mechanism. When it comes to permanent tax avoiding
strategies on the expense side, a few types of expenses are not tax
deductable. These expenses include nes, penalties, and bribes. As
for temporary tax-avoiding strategies, Finnish rms can also make
use of a depreciation reserve and depreciation adjustment (see for
example Niskanen & Keloharju, 2000). When companies make
investments, they decide on a planned schedule for depreciations.
Every year, they can then decide (within the limits allowed in the
tax laws) to depreciate more or less than planned. If they
depreciate less, they accumulate tax reserves, which are reported
in the balance sheet. This can then be used in later years to reduce
the amount of prots and the amount of taxes paid. An additional
way to avoid tax payments in open European economies such as
Finland is to set up subsidiaries in countries with lower tax rates or
to channel some of the operations through countries with lower
tax rates.1
1
Recent evidence suggests that for example a Finnish-Swedish Pulp- and Paper
company StoraEnso has avoided approximately 50 million EUR in taxes by
channeling its pulp-sales through the Netherlands (Finer, Laine, & Ylonen, 2012).

So, contrary to low tax alignment countries such as the US, tax
aggressive behaviour becomes visible in the nancial statements of
rms in high tax alignment countries. Consequently, tax aggressive
behaviour has a real impact for rms in high alignment countries:
the rms real economic performance may not become visible in
their nancial reports due to tax aggressive behaviour. This may
make it very difcult for shareholders and other stakeholders to
understand and value the true economic performance of the rm.
Therefore, studying the determinants of tax aggressiveness in the
context of high tax alignment countries is very important.
Our article contributes to the literature in several ways. First,
our study is the only study focusing on tax aggressiveness in a
private family rm context. Prior research generally focuses on
differences in rms tax reporting between private and public rms
(e.g. Beatty & Harris, 1999; Mills & Newberry, 2001) or between
public family rms versus nonfamily rms (for example Chen et al.,
2010). We focus only on private (family) rms because specic
agency problems in private family rms make us eager to believe
that there are different agency problems within the heterogeneous
group of private family rms leading to differences in tax
aggressiveness. Additionally, we take into account the socioemotional wealth perspective which complements the agency
view. Moreover, previous studies only investigate the direct effect
of board monitoring on tax aggressive behaviour (e.g. Lanis &
Richardson, 2011; Minnick & Noga, 2010). In this article, we study
the moderating effect of board monitoring, which can shed a new
light on this stream of literature. Additionally, the existing
literature on tax aggressiveness is dominantly US based (which
is a low tax alignment country) and does not necessarily translate
to other high tax alignment countries such as Finland.
This article proceeds as follows. In the next section, the
theoretical underpinnings are discussed and hypotheses are
derived. In section Data and variables, the dataset and variables
are discussed. Section Results presents our results and section
Discussion and conclusion highlights the major conclusions and
implications.
Literature review and hypotheses development
According to traditional agency theory, the privately, family
owned and managed rms are often considered as a low agency
cost case (Fama & Jensen, 1983; Jensen & Meckling, 1976). Family
members would be more likely to behave altruistically. Parental
altruism is a utility function in which the welfare of parents is
positively linked to the welfare of their children. Altruism may
have several benecial effects such as the creation of a selfreinforcing system of incentives encouraging family members to
be considerate of one another (Schulze, Lubatkin, & Dino, 2003a)
and the enforcement of incentives to communicate and cooperate
with each other (Van den Berghe & Carchon, 2003). When a rm is
owned solely by a single owner-manager, it can even be considered
as a zero agency cost case (Ang, Cole, & Lin, 2000).
However, by (partially) separating ownership from management in private family rms, agency costs may arise due to
information asymmetries and strains on the limits of bounded
rationality among family owners. The interests of owner(s) and
manager(s) may not be completely aligned: the ability of the CEO
to act in his own interests at the expense of (other) family rm
owners will increase (Chua, Chrisman, & Sharma, 2003). Engaging
in tax aggressive behaviour by the CEO may be a reection of this
shareholder-manager agency problem (Hanlon & Heitzman, 2010).
Engaging in tax aggressive activities is accompanied by costs
and benets within the context of private family rms. As Dyreng
et al. (2010) indicate that the CEO plays an economically signicant
role in determining the level of tax avoidance that rms undertake,
we take the perspective of the CEO in studying the costs and

T. Steijvers, M. Niskanen / Journal of Family Business Strategy 5 (2014) 347357

benets of tax aggressive behaviour that determine the actual

extent of tax aggressiveness. CEOs are generally not tax experts but
they set the tone at the top, which may explain their role in a rms
level of tax aggressiveness (Huseynov & Klamm, 2012).
Chen et al. (2010) provide an overview of these costs and
benets in public rms. On the benet side, the direct tax savings
are included which benet all shareholders. Moreover, the
complexity and obscure nature of tax aggressiveness may allow
the CEO to mask or hide any kind of rent extraction activities (for
example perk consumptions and excessive compensation). On the
cost side, the rm risks a potential penalty by the tax authorities.
Moreover, if other shareholders perceive tax aggressive behaviour
as a way to mask rent extraction, a price discount will be imposed
on the rms shares.
However, with respect to private family rms, we argue that the
costs may differ from those of public rms. First, contrary to public
rms (Chen et al., 2010), this rent extraction and other perquisite
consumption behaviour by the CEO (Anderson & Reeb, 2003;
Schulze et al., 2003a), will not be punished by the shareholders by a
price discount of the shares because private ownership lacks
disciplining of the market for corporate control. Moreover, the lack
of external discipline increases the likelihood that information
asymmetries will develop vis-a`-vis, for example, outside shareholders (Lubatkin, Schulze, Ling, & Dino, 2005). Additionally,
previous studies indicate that CEO turnover is signicantly lower
in family rms, indicating that possible rent extraction is less likely
to be punished by the shareholders (Tsai, Hung, Kuo, & Kuo, 2006).
Secondly, since private family rm owners are underdiversied
and have their wealth tied disproportionately to their rms
(Anderson, Mansi, & Reeb, 2003), any penalty from the tax
authorities is more likely to be substantial to them. Thirdly, there is
the possible damage caused to the CEOs reputation but especially
the rms reputation (Dyer & Whetten, 2006). Due to the large
equity ownership by the family, private family rms have a much
longer investment horizon and greater reputation concerns
compared to public rms or private nonfamily rms (Gedajlovic
& Carney, 2010). They want to pass the rm onto the heirs and
want to preserve the reputation of the family name (Berrone, Cruz,
& Gomez-Mejia, 2012).
Therefore, by taking a socioemotional wealth perspective, we
argue, based on Berrone et al. (2012) that family owners will frame
problems and their decision making in terms of how the resulting
actions will affect socioemotional wealth. Socioemotional wealth
(hereafter SEW) refers to the non-nancial aspects of the rm that
meet the familys affective needs, such as identity, the ability to
exercise family inuence, and the perpetuation of family dynasty
(Gomez-Mejia et al., 2007, p. 106). This SEW model is in line with
the proposition by Maciejovsky, Schwarzenberger, and Kirchler
(2012) stating that the specic decision to engage in tax aggressive
behaviour cannot be purely explained by economic variables.
Family rms have been represented as a combination of two
systems that overlap and interact: an emotion-oriented family
system focussing on noneconomic goals and the results-oriented
business system focussing on economic goals (Classen, Van Gils,
Bammens, & Carree, 2012; Gersick, Davis, Hampton, & Lansberg,
1997). Therefore, family rms do not only have nancial goals (for
example reducing taxes) but also noneconomic goals (Chrisman
et al., 2010), which help explain why family rms behave
distinctively. Or as Berrone et al. (2012, p. 2) state: gains or
losses in SEW represent the pivotal frame of reference that familycontrolled rms use to make . . . policy decisions.
So, preserving the socioemotional wealth is itself a key goal in
many private family rms (Stockmans, Lybaert, & Voordeckers,
2010). For example, Berrone et al. (2010) found that family
controlled rms in polluting industries tend to contaminate less in
order to enhance the familys image and protect SEW. Moreover,

349

empirical evidence is available that suggests that because of strong

identication with the rms name and because public condemnation could be emotionally devastating for the family, family
rms exhibit higher levels of community citizenship (Berrone
et al., 2010) and take care to perpetuate the positive family image
and reputation (Sharma & Manikuti, 2005).
Therefore, we argue that the CEOs behaviour in a family rm
with respect to tax aggressiveness cannot be understood without
taking into account noneconomic emotional aspects captured by
the SEW perspective. Thus, we argue that, in general, private family
and nonfamily rms can cope with agency problems where the
CEO can react by engaging in tax aggressive behaviour. However,
we hypothesize, based on the SEW perspective, that the component of commitment to the preservation of SEW, specic for
(private) family rms, will outweigh the (agency) benets of tax
aggressive behaviour in private family rms. We therefore propose
the following hypothesis:
Hypothesis 1. Private family rms exhibit a lower level of tax
aggressive behaviour compared to private nonfamily rms.
As private family rms cannot be viewed as a homogeneous
entity (Chrisman, Chua, & Sharma, 2005; Westhead & Howorth,
2007), specic family rm characteristics may inuence their
agency problems as well as their tendency to preserve SEW and
resulting tax aggressive behaviour. Given the importance of the
CEO as a key driver of corporate behaviour in this context (Dyreng
et al., 2010) and the effect of the extent of separation between
ownership and control on agency costs and the resulting tax
aggressive behaviour, we take into account the CEOs ownership
share.
If the CEO has a high ownership share, agency costs are expected
to be low. Agency theory expects the agency costs to decrease
when the CEOs ownership share increases. More specically, it is
assumed that the more shares he has, the less he will be inclined
towards consuming perquisites to maximize their own utility as
the fraction of the costs the CEO has to bear for consuming these
perquisites is positively related with the percentage of ownership
(Jensen & Meckling, 1976). So, in that case, the CEO bears many of
the costs and receives nearly all of the benets of any of his actions
including tax aggressiveness (Fama & Jensen, 1983).
Moreover, referring to the SEW perspective, a CEO with a high
ownership share is mainly worried by passing the rm to his
children (Blanco-Mazagatos, de Quevedo-Puente, & Castrillo,
2007). According to Gomez-Mejia et al. (2007), a strong family
CEO increases the focus on family goals and the persistence to
protect SEW. They may be less eager to engage in rent extraction
because this may harm the rm. Parental altruism gives the
controlling owner/CEO incentive to take actions that they believe
would benet the nuclear family (Lubatkin et al., 2005; BlancoMazagatos et al., 2007). They tend to focus on family goals at the
expense of other nancial goals (Westhead, 2003). In addition, the
emotional attachment to the rm and the self identication with
the rm are strong leading them to do the right thing for the rm
and the family (Gomez-Mejia et al., 2007; Schulze et al., 2003a).
Therefore, we argue that the CEO will be less inclined to engage in
tax aggressive activities because not only the avoidance of any
penalty from the tax authorities is important, but also the negative
publicity or loss of SEW are essential.
A CEO with a lower ownership share, usually arising due to
succession of the rm over several generations, may be more
inclined to engage in tax aggressive activities. First, agency costs
are expected to increase. Family ties and altruistic feelings weaken
and the family CEO with a low ownership share will often put the
welfare of the own nuclear family before the wealth of the
extended family (Karra, Tracey, & Phillips, 2006; Lubatkin et al.,
2005). This low ownership share may reduce the motivation of

350

T. Steijvers, M. Niskanen / Journal of Family Business Strategy 5 (2014) 347357

descendant CEOs, which increases the incentive to act opportunistically because they bear only part of the cost of such an action.
It may enhance rent extraction by the CEO leading to, for example,
increased perquisite consumption and additional remuneration
(Fama & Jensen, 1983). The shareholder-manager agency conict
becomes more prominent.
Moreover, the low CEO ownership share weakens the attachment of the family to the rm. The utility generated by the
preservation of SEW decreases as the rm moves into later
generational stages (Gomez-Mejia et al., 2007). Throughout
generations, the focus shifts from family goals to a combination
of family and business goals (Schulze, Lubatkin, & Dino, 2003b).
Therefore, the rm reputation effect of tax aggressive behaviour
and the incentive to preserve SEW turn to be of minor importance
since the family ties have weakened and the intra family conict
may intensify (Miller & Le Breton-Miller, 2006).
If the CEO has no ownership share and is thus a professional
nonfamily CEO, ownership and management are completely
separated which may lead to signicant shareholder-manager
agency costs due to misalignment of incentives. Goals of manager
(agent) and owner(s) (principal) can diverge because the
nonfamily manager is not always familiar with the family goals
or may choose other goals than those strived for by the family
shareholders (Jensen & Meckling, 1976). He will have a more short
term view compared to a family CEO (Miller & Le Breton-Miller,
2006). He will be more inclined to improve the nancial results for
the family shareholders and engage in tax aggressive activities,
such as setting up group structures abroad. Since family rms are
not eager to provide nonfamily managers with equity shares, they
will be more likely to receive a salary or bonus based on their
performance (Banghoj, Gabrielsen, Petersen, & Plenborg, 2010).
Additionally, he may increase the free cash ow by engaging in tax
aggressive behaviour by making use of depreciation reserves and
adjustments, in order to invest in pet projects or pursue personal
objectives (Jensen, 1986).
As nonfamily CEOs shift a rms orientation towards short term
nancial goals (Classen et al., 2012), we expect them to be less
concerned with penalties from the tax authorities or other long
term implications with regard to rm reputation or SEW. In line
with this reasoning, Gomez-Mejia et al. (2007) and Gersick,
Lansberg, Desjardins, and Dunn (1999) state that nonfamily CEOs
will pursue the SEW objectives less than family CEOs because
personal attachment and self-identication with the rm is less
strong in nonfamily managed rms. Nonfamily CEOs are brought in
to provide objectivity and more rationality (Blumentritt, Keyt, &
Astrachan, 2007). Their relationship with the rm is more distant,
transitory and individualistic (Block, 2011). Therefore, only the
potential personal reputation damage may withhold the nonfamily
CEO from engaging in excessive tax aggressive behaviour. Contrary
to a family CEO who holds a rather secure position in the rm,
nonfamily CEOs can be laid off more easily, making them more
concerned about their personal reputation on the market for
corporate executives (Allen & Panian, 1982; Block, 2010). Based on
all these arguments, we posit:
Hypothesis 2. Private family rms with a high CEO ownership
stake exhibit a lower level of tax aggressive behaviour.
However, the board of directors may in several ways be an
instrument to reduce shareholder-manager agency problems and
restrict tax aggressive behaviour by the CEO. The link between the
board of directors and tax aggressive behaviour is grounded in the
agency view of corporate governance. According to this view, rm
decision-makers may behave opportunistically by pursuing their
own interest (Fama & Jensen, 1983; Jensen & Meckling, 1976).
Therefore, agency theorists point at the instalment of a board of
directors as an instrument to control the rms decision-makers.

A board of directors provides advice, counselling and networking,

but also serves to align the interests of managers with shareholders
interests so as to safeguard shareholders interests (Johannisson &
Huse, 2000). The board of directors is legally responsible for
monitoring and evaluating senior management for the benet of
the rm (Forbes & Milliken, 1999).
So, within an effective corporate governance structure, the
board of directors must verify whether the rms management acts
in the best interest of the family and/or nonfamily shareholders. In
case of sound corporate governance, the directors should detect
any kind of rent extraction behaviour and report it to the
shareholders. In order to perform this task effectively, the directors
should have the necessary expertise and objectivity that ostensibly
mitigates the expropriation of rm resources, for example, by rent
extraction (Bammens, Voordeckers, & Van Gils, 2011). Therefore,
we argue that as rent extraction possibilities are reduced, the
incentive for a CEO to engage in tax aggressive behaviour to mask
rent extraction would be reduced. As argued above, private family
rms with a lower CEO ownership share would be more eager to
engage in tax aggressive behaviour. In those rms, the advantages
of tax aggressiveness become dominant since the preservation of
the rms reputation and SEW are less important due to weaker
family ties. However, the control role performed by an effective
board of directors will reduce the CEOs incentive to engage in tax
aggressive behaviour. The board has to avoid rent extraction such
as excessive CEO compensation and thereby reduces the motivation of the CEO to engage in tax aggressive activities. Moreover, in
case of tax related law suits, the board may be legally liable and
their reputation capital may also be threatened (Carcello,
Hermanson, Neal, & Riley, 2002). It may subject the directors to
heavy criticism. Therefore, we expect that they will monitor the
rms management to avoid that CEOs with low ownership share
engage in tax aggressive behaviour.
Therefore, we consider the moderating effects of an effective
board of directors with respect to its monitoring role, on the
relationship between the CEOs ownership share and tax aggressive
behaviour. In literature, several indicators for effective board
monitoring are suggested. First, the presence of outside board
members can signal effective monitoring by the board of directors.
From a traditional agency perspective, boards should be able to act
independent of those parties they are supposed to control. As a
result, their monitoring effectiveness increases, thereby decreasing
managerial opportunism (Harford, Mansi, & Maxwell, 2008).
Secondly, CEO duality, indicating that the CEO is also the chairman
of the board, is a mechanism that provides the CEO a considerable
concentration of power. Only a separation of both functions ensures
that the CEO has no unregulated power (Fama & Jensen, 1983;
Jensen, 1993). Therefore, CEO duality reduces the independence of
the board. Thirdly, board size can also be considered as a corporate
governance mechanism. A director from a large board may nd that
the costs of speaking out against top management may outweigh the
benets of carrying out his monitoring duty diligently (Bliss, 2011).
Larger boards may be more easily controlled/governed by a
dominant CEO and preserve the power of the CEO (Jensen, 1993).
Moreover, larger boards can also be characterized by free riding and
social loang problems (Dalton, Daily, Johnson, & Ellstrand, 1999).
Directors may be eager to exert less effort when they work in a large
board than when they work in a small group where their
contribution is more visible. So, we hypothesize:
Hypothesis 3a. The negative relationship between CEO ownership
and tax aggressive behaviour will be weakened by the presence of
outside board members.
Hypothesis 3b. The negative relationship between CEO ownership
and tax aggressive behaviour will be reinforced by the presence of
CEO duality.

T. Steijvers, M. Niskanen / Journal of Family Business Strategy 5 (2014) 347357

Hypothesis 3c. The negative relationship between CEO ownership

and tax aggressive behaviour will be weakened if there is a small
board of directors.
Data and variables
Data set and method
The data for this study were collected through a private survey.
Of the 3262 questionnaires sent, a total of 621 responses were
received, which resulted in an effective response rate of
19 percent. The nal sample consists of 600 SMEs operating in
Finland, because we drop rms which do not correspond to the EU
denition of SMEs. The rms represent all industries, excluding
primary production. The sample rms are rms with at least two
employees and whose legal form is a limited liability. The rms
were asked to provide information on their ownership structure
during the years 20002005, for each year separately. The rms
were also asked to provide information on their board composition during the time period. Non-respondent tests have been
performed for the database, and they suggest that the rms that
responded to the survey are statistically signicantly similar to
the ones that did not respond. The nancial data were collected
from the Voitto+ register. This register includes data on rm age,
employment, line of business, and the complete nancial
statements. A private family rm is dened as a rm where more
than 50 percent of the shares are owned by the family. This
denition is in line with the majority of family business
denitions that require family ownership as one of the main
indicators for dening a family rm (Chua, Chrisman, & Sharma,
1999).
After elimination of outliers and missing values, we ended up
with a nal unbalanced panel dataset of 1650 private family and
nonfamily rm-year observations including 898 family rm-year
observations. Each of the models will be estimated based on robust
OLS estimations. As indicated by Antonakis, Bendahan, Jacquart,
and Lalive (2010), standard errors have to be consistent in order to
obtain valid results. However, heteroscedasticity of the residuals
are a threat to this validity, leading to inconsistent standard errors.
If standard errors are not correctly estimated, the p-values of the
beta-coefcients will be over- or understated. In order to check for
heteroskedasticity, we performed the White test, which indicated
a problem of heteroskedasticity of the residuals (x2 = 22.92; pvalue = 0.00001). Therefore, we use robust (consistent) standard
errors in our OLS estimations.
Measures
The dependent variable we use is the effective tax rate
dened as total tax expense divided by earnings before taxes.
Firms that are more tax aggressive have lower effective tax rates
(ETRs). In a Finnish context, some of the tax aggressiveness
strategies (such as increasing depreciation reserves) achieve a
lower ETR through increasing accounting expenses, thereby
reducing taxable income, taxes and annual net income. Other
strategies (such as locating operations in low tax countries)
achieve a lower ETR through lower taxes and thereby increased
net income. So, what is common for tax aggressive strategies, is
that they reduce the effective tax rate. Therefore, we argue that
the ETR is a viable measure for our research setting where
detailed information on the tax aggressiveness choices made is
not easily available.
Moreover, this measure is widely used in previous research (e.g.
Chen et al., 2010; Chyz, Leung, Li, & Rui, 2013; Lanis & Richardson,
2011; Minnick & Noga, 2010). Even though previous studies

351

conducted in low tax alignment countries also make use of a

variety of other measures for tax aggressiveness derived from the
book-tax difference (e.g. Chen et al., 2010; Chyz et al., 2013; Desai
& Dharmapala, 2006), they cannot be used in this study because
Finland is a high tax alignment country where nancial statements
are taken as the basis for taxation. Consequently, there is book-tax
alignment.
Additionally, evidence from the US suggests that nancial
accounting income has become increasingly higher than taxable
income. Hanlon and Shevlin (2005) argue that this may be a result
of earnings management in nancial statement income and tax
aggressiveness in taxable income. This suggests that when the
results of tax planning and earnings management have to be
reported simultaneously (which is the case in Finland, as a high tax
alignment country), all the relevant information should be
captured by the ETR. However, this ETR will mainly capture the
tax aggressiveness because, as Burgstahler, Hail, and Leuz (2006)
suggest, there is less earnings management in countries with tax
alignment. Moreover, Hanlon and Heitzman (2010) report that the
potential benets of book-tax conformity include that management will be more truthful when reporting the income numbers
(less upward earnings management) because anything reported
will also be taxed.
We incorporate several independent variables in our study. In
order to verify whether family rms are more or less tax aggressive
than nonfamily rms, we incorporate the ownership percentage in
hands of the family (Familyown). Moreover, we can also
incorporate a dummy variable indicating whether the rm can
be considered as a family rm. Therefore, we incorporate a dummy
variable Family50 with a value 1 if more than 50 percent of the
shares are owned by the family; 0 otherwise. Alternatively, we
included a dummy variable Family0 with a value 1 if the rm
indicates that it is a family rm with some amount of the shares
(>0 percent) in hands of the family; 0 otherwise.
Within the group of private family rms, we include Ceoown
which measures the amount of shares owned by the CEO. We have
no information on whether the CEO is a family or a nonfamily
member. However, if CEO ownership (Ceoown) is 0, which is the
case in 17% of our sample, it will generally be someone who is not
part of the family.2 CEO duality (Ceo_dual) has a value 1 if the
rms CEO is also the chair of the board of directors; 0 otherwise.
The variable Ext has a value 1 if the board contains at least one
outside board member who does not belong to the management
team; 0 otherwise. The size of the board of directors (Board size)
consists of the number of directors.
In each of the regressions we perform, we control for rm
characteristics reported in prior literature (Chen et al., 2010; Frank
et al., 2009; Yuan, McIver, & Burrow, 2012) that are correlated with
tax aggressive behaviour. Therefore, we can ensure that our results
are not driven by fundamental differences between family and
nonfamily rms and within the group of family rms. In our study,
we control for the rms protability by incorporating the return
on assets (Roa), the rms leverage measured by long term debt
(Lev), plant, property and equipment (Ppe) and intangible
assets (Intang). In line with Chen et al. (2010), these control

2
If the family rm is inherited by a family member who becomes CEO of the rm,
(part of) the shares are transferred. However, when a nonfamily CEO is hired, he
generally obtains no ownership share (Michielsen, Voordeckers, Lybaert, &
Steijvers, 2013). Family rms are eager to transfer the rm to the next generations
and to keep control within the family in order to perpetuate the family dynasty.
Therefore, family rms are reluctant to hire nonfamily CEOs because the family
wants to avoid the loss of strategic and operational control and goal conicts
(Gomez-Mejia, Cruz, Berrone, & De Castro, 2011). However, even when hiring a
nonfamily CEO is necessary for the rm, providing him/her with shares is often
considered to be a step further and a step too far in the family losing control of the
rm.

T. Steijvers, M. Niskanen / Journal of Family Business Strategy 5 (2014) 347357

352

Table 1a
Descriptives and Pearson correlation matrix (total sample).

1.
2.
3.
4.
5.
6.
7.
8.
9.

ETR
Family50
Family0
Familyown
Roa
Lev
Ppe
Intang
Size (in 000)

Mean

Std. dev.

0.23
0.57
0.58
0.53
0.20
0.22
0.34
0.03
310

0.14
0.49
0.49
0.47
0.33
0.53
0.35
0.14
2.90

2
1
0.06**
0.06***
0.05**
0.24***
0.10***
0.06**
0.09***
0.08***

3
1
0.97***
0.96***
0.03
0.03
0.10***
0.10***
0.01

1
0.95***
0.03
0.03
0.10***
0.10***
0.03

1
0.02
0.03
0.12***
0.10***
0.01

1
0.09***
0.03
0.04*
0.13***

1
0.35***
0.19***
0.05**

1
0.05**
0.01

1
0.08***

10

1
0.27***

N = 1650.
*
Signicant at the 10 percent level (two-tailed test).
**
Signicant at the 5 percent level (two-tailed test).
***
Signicant at the 1 percent level (two-tailed test).

Table 1b
Descriptives and Pearson correlation matrix (subsample of family rms).

1. ETR
2. Ceoown
3. Size (in 000)
4. Roa
5. Lev
6. Ppe
7. Intang
8. Ceo_dual
9. Ext
10. Board size

Mean

Std. dev.

0.23
0.50
315
0.21
0.22
0.37
0.02
0.53
0.81
2.35

0.14
0.33
3.20
0.21
0.34
0.39
0.04
0.50
0.39
0.97

2
1
0.06*
0.03
0.25***
0.16***
0.13***
0.05*
0.07**
0.02
0.06*

3
1
0.22***
0.06*
0.03
0.03
0.04
0.35***
0.01
0.33***

1
0.17***
0.05
0.03
0.01
0.15***
0.03
0.37***

1
0.22***
0.03
0.01
0.04
0.07**
0.10***

1
0.60***
0.06*
0.03
0.03
0.01

1
0.02
0.06*
0.10***
0.04

1
0.03
0.08**
0.04

1
0.05
0.34***

N = 898.
*
Signicant at the 10 percent level (two-tailed test).
**
Signicant at the 5 percent level (two-tailed test).
***
Signicant at the 1 percent level (two-tailed test).

variables were scaled by lagged total assets.3 For rms with a higher
protability, it can be expected that they tend to have higher
effective tax rates (ETR). Firms with a higher leverage (Lev) will
have more interest costs and thus lower ETR. For plant, property and
equipment (Ppe) and intangible assets (Intang), its depreciations
are tax deductible. So we can expect a positive relation between the
extent of Ppe and Intang and the resulting depreciation on the
one hand and the ETR on the other hand. Lastly, we also control for
rm size by including the natural logarithm of total assets of the
previous year (Size). In addition, for all regressions, we include
dummies to control for year and industry xed effects.
Tables 1a and 1b present the descriptive statistics and Pearson
correlation matrix for the main variables of our analyses for the total
sample of private family and nonfamily rms and for the subsample
of private family rms, respectively. The correlation tables indicate
no problem of multicollinearity. Moreover, we computed the
variance ination factor analysis (VIF) among the variables
indicating how the variance of an estimator is inated by the
presence of multicollinearity (Gujarati, 1995). The highest VIF value
here is 7.55 for the subsample of family rms and 1.73 for total
sample of family and nonfamily rms, which is below the threshold
of 10, so no multicollinearity is present (Manseld & Helms, 1982).
Table 1a shows that more than 50% of the total sample consists
of family rms. Table 1b reveals that the private family rms in our
sample have an average asset size of 315,000 euros and have an
average effective tax rate (ETR) of 23.3 percent. The median ETR for
3
Lagged assets have not been affected by current years decisions. For example,
for ROA, the assets of year t 1 (also beginning assets of the current year t) are used
to create the prots in year t. The lagged assets are also preferred because the
current year assets would include the prots. The same reasoning can be applied to
the other control variables. We also re-estimated the base regression (Table 3,
model (1)) using assets in year t as a denominator but the results remain
qualitatively the same.

the family rms in our sample is 28%. Therefore, we can conclude

that the ETR of the rms in our sample is in line with the general
corporate tax rate for Finnish rms (29% until 2004; reduced to 26%
in 2005). On average, the CEO owns 50 percent of the shares. The
rms are characterized by a rather high ROA of 20.9 percent.
Moreover, in more than 50 percent of the rms, the CEO is also the
chair of the board of directors. In the majority of rms
(81.2 percent), an outside director is serving on the board. On
average, 2.3 board members serve on the board.
Results
In Tables 2 and 3, the results are presented. All regression
models are estimated with Ordinary Least Squares (OLS) and
Table 2
Robust OLS regression on SMEs tax aggressiveness.
Dep. variable: ETR

(1)

(2)

(3)

**

Family50
0.01 (0.01)
Family0
0.01** (0.06)
Familyown
0.01* (0.01)
Roa
0.11*** (0.03)
0.11*** (0.03)
0.11*** (0.03)
Lev
0.01 (0.01)
0.01 (0.01)
0.01 (0.01)
Ppe
0.03** (0.01)
0.03** (0.01)
0.03** (0.01)
Intang
0.09 (0.08)
0.09 (0.08)
0.09 (0.08)
Size
0.01*** (0.01)
0.01*** (0.01)
0.01*** (0.01)
Constant
0.13*** (0.02)
0.12*** (0.02)
0.13*** (0.02)
R2
0.11
0.11
0.11
Adjusted R2
0.10
0.10
0.10
***
***
F value
8.69
8.58
8.43***
Number of observations 1650
1650
1650
Note: Robust asymptotic standard errors are reported in parentheses.
*
Signicant at the 10 percent level (two-tailed test).
**
Signicant at the 5 percent level (two-tailed test).
***
Signicant at the 1 percent level (two-tailed test).

T. Steijvers, M. Niskanen / Journal of Family Business Strategy 5 (2014) 347357

353

Table 3
Robust OLS regression on family rms tax aggressiveness.
Dep. variable: ETR
Ceoown
Ceoown  Size
Ext
Ceoown  Ext
Ceo_dual
Ceoown  Ceo_dual
Board size
Ceoown  Board size
Roa
Lev
Ppe
Intang
Size
Constant
R2
Adjusted R2
Change in R2
F value
Number of obs.

(1)

(2)

(3)
**

0.02 (0.01)

0.15 (0.07)
0.03*** (0.01)

0.13

(4)
***

(0.03)

(5)
0.05* (0.03)

0.01 (0.02)

0.06*** (0.02)
0.13*** (0.03)
0.01 (0.02)
0.01 (0.03)

***

0.17 (0.03)
0.01 (0.04)
0.04 (0.03)
0.16 (0.13)
0.01*** (0.01)
0.15*** (0.03)
0.12
0.10
0.12***
8.48***
898

***

***

0.17 (0.02)
0.01 (0.05)
0.04 (0.03)
0.16 (0.13)
0.01 (0.01)
0.22*** (0.04)
0.13
0.11
0.01***
8.77***
898

0.01 (0.01)
0.02 (0.01)
0.17*** (0.02)
0.01 (0.04)
0.05 (0.03)
0.15 (0.13)
0.01***(0.01)
0.14***(0.04)
0.13
0.10
0.01
7.58***
898

***

0.17 (0.02)
0.01 (0.04)
0.04 (0.03)
0.17 (0.13)
0.01*** (0.01)
0.11*** (0.04)
0.14
0.12
0.02***
8.52***
898

0.17 (0.02)
0.01 (0.04)
0.04 (0.03)
0.15 (0.13)
0.01*** (0.01)
0.15*** (0.04)
0.13
0.10
0.01
8.01***
898

Note: Robust asymptotic standard errors reported in parentheses.

* Signicant at the 10 percent level (two-tailed test).
**
Signicant at the 5 percent level (two-tailed test).
***
Signicant at the 1 percent level (two-tailed test).

ETR as rm size changes, indicated by the solid line. The dotted

lines represent the 95 percent condence interval, which allows us
to determine the conditions under which the ownership share of
the CEO has a signicant impact on the rms ETR. The effect is
only signicant if both the upper and lower bounds of the
condence interval are above or below the zero line.
Fig. 1 shows that the ownership share of the CEO has a
signicant positive effect on ETR, indicating a lower extent of tax
aggressiveness if the family rm has more than 400,000 euro of
assets (at the point where ln(size) equals 6). The positive effect
increases as rm size increases. This conrms our hypothesis
2 indicating that as the ownership share of the CEO increases, the
family rm becomes less tax aggressive. For the smaller family
rms in our database, we nd no signicant effect of Ceoown on
the tax aggressive behaviour. The signicant negative effect
revealed in Fig. 1, is related to rms with a rm size of less than
12,000 euro of assets, which are not present in our sample.
Additionally, we performed a constrained estimation to verify that
there is a net effect of Ceoown (negative) and CeoownxSize

[(Fig._1)TD$IG]
Marginal Effect of 'Ceoown' on tax aggressive behavior
Dependent Variable: ETR

Marginal Effect of Ceoown

.1 .2
-.2 -.1 0

robust standard errors are calculated. Table 2 reveals that

Hypothesis 1 can be conrmed. Private family rms appear to
be less tax aggressive compared to nonfamily rms. The dummy
variables Family50 and Family0 as well as Family own reveal
a signicant positive effect. The other signicant control variables
have the expected sign. More protable (Roa) and larger rms
(Size) seem to have a larger ETR whereas rms with more plant,
property and equipment (Ppe) have a lower ETR.
However, as argued above, private family rms are a
heterogeneous group. Therefore, further analysis within the group
of private family rms is provided in Table 3.
Regression (1) in Table 3 does not seem to conrm Hypothesis 2,
with respect to the effect of the CEO ownership share on tax
aggressive behaviour. Regression (1) shows no signicant effect of
Ceoown. However, rm size may be an important moderator in
the context of tax aggressive behaviour. Very small, young
private family rms may not have the experience to engage in
tax aggressive behaviour and are fully occupied with the core
business and/or survival of the rm. Therefore, we included in
regression (3) the moderating effect of rm size (Size) on the
relationship between Ceoown and ETR. We included the term
Ceoown  Size.
From the results in Table 3, we cannot deduct from regression
(2) what the impact is of the ownership share of the CEO on tax
aggressive behaviour when rm size changes.4 In order to capture
the total effect, we have to take into account the coefcient of
Ceoown and of the interaction term and the value of the
moderating variable which is Size (Brambor et al., 2006; Kam &
Franzese, 2007; Steijvers & Niskanen, 2013). Fig. 1 graphically
presents the marginal effect of the ownership share of the CEO on

1
4

Several econometrical studies (Berry, DeMeritt, & Esarey, 2010; Brambor, Clark,
& Golder, 2006; Norton, Wang, & Ai 2004) state that the effect of any independent
variable X in an interactive model with a continuous moderator Z, on the dependent
variable Y is not any single constant. The effect depends on the betas of X and of the
interaction term XZ, as well as on the value of Z, the moderating variable. In order to
interpret the results, the calculation of marginal effects is of great importance as it is
perfectly possible that these effects are signicant for relevant values of the
moderating variable, even if the coefcient on the interaction term is insignicant
(Berry et al., 2010; Brambor et al., 2006). More specic, we take into account the
relevant elements of the variance-covariance matrix and recalculate the standard
errors as suggested by Brambor et al. (2006, p. 74).

5
7
size (ln (assets))

11

Marginal Effect of Ceoown

95% Confidence Interval

Fig. 1. Marginal effect of CEO ownership share on tax aggressive behaviour:

extension of regression (2) in Table 3. This gure graphically presents the marginal
effect of CEO ownership share as rm size changes, indicated by the solid line. The
dotted lines represent the 95% condence interval, which allows us to determine
the conditions under which CEO ownership share has a signicant impact on the
rms tax aggressive behaviour. The effect is only signicant if the upper and lower
bounds of the condence interval are both above or below the zero line.

T. Steijvers, M. Niskanen / Journal of Family Business Strategy 5 (2014) 347357

354

(positive) together. We compare model (2) which is the unconstrained model, with a constrained model where we add a
constraint that Ceoown and the interaction term with size
together do not lead to a signicant effect. Based on the Wald test,
we obtain a signicant F value (F(1, 877) = 4.78; p = 0.03) which
means we can reject the hypothesis that there is no net effect.
Alternatively, we performed an F test, comparing the constrained
with the unconstrained model (F(1, 877) = 4.81; p = 0.03) which
again means we can reject that there is no net effect.
Furthermore, we study in regression (3), (4) and (5) of Table 3
the moderating effect of the board of directors performing
adequately their monitoring role. With respect to Hypothesis 3a
on the presence of outside board members, the variable Ext is
included in regression (3) as well as the interaction term
Ceoown  Ext. The interaction term has the predicted negative
sign and is signicant at 1 percent level. The signicant coefcient
of Ceoown equals 0.13, whereas the interaction term has a
coefcient of 0.13. This means that if family rms hire an external
board member, the effect of the ownership share of the CEO
(Ceoown) is reduced to 0. This indicates that the CEO ownership
share no longer affects the tax aggressive behaviour of the rm if
the board includes an independent outside director. The benet for
CEOs to engage in tax aggressive behaviour, at the expense of other
shareholders, is nearly absent when the rm hires an external
board member due to a higher monitoring effectiveness thereby
limiting possible rent extraction behaviour. This conrms Hypothesis 3a.
Regression (4) extends this discussion and takes into account
the moderating effect of CEO duality (CEO_dual). Therefore, we
include in regression (4), Ceo_dual and the interaction term
Ceoown  Ceo_dual. Contrary to what we expected, CEO duality
appears to have no signicant moderating effect. It seems that
whether the board is chaired by the CEO or not, it does not
inuence the effect of CEO ownership share on tax aggressive
behaviour. Therefore, we cannot conrm Hypothesis 3b. Finally, in
regression (5), we include the variable Board size as well as the
interaction term Ceoown  Boardsize. Again, in line with
regression (2), in order to capture the effect of the CEOs ownership
share when board size changes, we have to take into account the
marginal effect of the CEOs ownership share. These marginal
effects are drawn in Fig. 2.
However, Fig. 2 reveals no signicant moderating effects for any
value of board size. Thus, we cannot conrm hypothesis 3c. In

[(Fig._2)TD$IG]
Marginal Effect of 'Ceoown' on tax aggressive behavior

-.2

Marginal Effect of Ceoown

.1
-.1
0
.2

Dependent Variable: ETR

4
Board size

Marginal Effect of Ceoown

95% Confidence Interval

Fig. 2. Marginal effect of CEO ownership share on tax aggressive behaviour:

extension of regression (5) in Table 3. This gure graphically presents the marginal
effect of CEO ownership share as board size changes, indicated by the solid line. The
dotted lines represent the 95% condence interval, which allow us to determine the
conditions under which CEO ownership share has a signicant impact on the rms
tax aggressive behaviour. The effect is only signicant if the upper and lower
bounds of the condence interval are both above or below the zero line.

conclusion, with respect to the monitoring effectiveness of the

board, only the presence of external board members seems to be
effective to reduce the incentives for CEOs to engage in tax
aggressive behaviour. Separation of the positions of chairman of
the board and CEO as well as board size do not play a moderating
role. These insignicant results may be explained by the lack of an
active board of directors. Smaller boards or a separation of the
functions of chairman and CEO can only be effective governance
and monitoring tools if the board is not a rubber stamp board.
Moreover, the descriptive statistics indicate that the average board
size is rather small with 2.3 board members which may suggest
that many boards may not be active boards. However, results seem
to indicate that the presence of an outside board member may
ensure the activeness of the board, performing adequately their
monitoring role.
In order to verify the robustness of our results, we additionally
used the difference between the nominal and reported tax rate as
dependent variable (results not reported). Regression results are
comparable to the results presented in Tables 2 and 3. We only
notice one slight difference with respect to model 1 in Table 3,
based on a subsample of family rms. Using the difference between
the nominal and reported tax rate as dependent variable,
Ceoown reveals a positive signicant effect (only at 10%
signicance level) on the dependent variable. This positive
signicant effect is again in line with Hypothesis 1. When using
ETR as dependent variable, this positive signicant effect was only
found for larger rms, however at a 5% signicance level.
Discussion and conclusion
Contributions
Corporate tax aggressiveness is a very young but active research
domain. Even though this topic has received some attention in the
academic literature, these studies only focus on low tax alignment
countries. However, tax aggressive behaviour is also important to
study in a high tax alignment country such as Finland. Contrary to
low tax alignment countries, tax aggressive behaviour becomes
visible in the nancial statements of rms in high tax aligned
countries which makes it difcult for stakeholders to value the true
economic performance of the rm. Therefore, studying the
determinants of tax aggressiveness in the context of high tax
alignment countries is even more important. With our contribution, we want to advance the knowledge on the determinants of tax
aggressiveness by using a principal-agent setting, more specically
in the context of private family rms.
In addition, in this article, we argue that tax aggressive
behaviour in private family rms is a concept that is too complex
to be explained only by economic drivers. Therefore, we
complement the agency view with a socioemotional wealth
(SEW) perspective. As indicated by Berrone et al. (2012, p. 5),
the SEW model naturally stems from the reality of family
businesses that suggest the existence of multiple salient goals that
are driven by the values of the family and that change over time.
Therefore, SEW is a key noneconomic reference point in family
rms for deciding to engage in a certain extent of tax aggressive
behaviour. In this article, we examine the extent of tax aggressiveness of private family rms, relative to their nonfamily
counterparts. We nd that private family rms appear to be less
tax aggressive than private nonfamily rms. This result highlights
the importance of the noneconomic costs related to tax aggressive
behaviour being the possible rm reputation damage and loss of
SEW. Even though tax aggressive behaviour provides tax savings
and allows the CEO to mask rent extraction to the detriment of
other shareholders, the economic costs seem to outweigh the
benets.

T. Steijvers, M. Niskanen / Journal of Family Business Strategy 5 (2014) 347357

Within the group of private family rms, our article contributes

to a better understanding of the impact of the ownership structure
on private family rms tax reporting. Results seem to indicate that
private family rms with a higher CEO ownership stake are less
eager to engage in tax aggressive behaviour, while CEOs with a
lower or no ownership share are more eager to engage in tax
aggressive behaviour. This result highlights the importance of the
unique agency conict between the CEO (agent and possibly
principal) and (other) shareholders (principals) in determining
private family rms tax aggressive behaviour. Additionally, results
seem to conrm the role of SEW in explaining the CEOs tax
aggressiveness.
Alternatively, one might interpret these results differently. Tax
aggressive behaviour by CEOs with low or no ownership share,
being professional CEOs, might simply be an indication of good
professional cash ow management. However, nonfamily or
professional CEOs cannot as such be equated to being managers
that engage in good professional management while family CEOs
are nonprofessional. Or as Dekker, Lybaert, Steijvers, Depaire, and
Mercken (2012) argue: Moreover, the sense of equating professional managers with external, nonfamily managers, leads to the
outdated assumption that family members are inherently nonprofessional managers (e.g. Bennedsen, Nielsen, Perez-Gonzalez, &
Wolfenzon, 2007; Berenbeim, 1990; Bloom & Van Reenen, 2007).
Both family and nonfamily managers can possess specialized
technical knowledge (Corbetta, 1995) and formal business training
(Dyer, 1988) in order to engage in professional cash ow
management. Moreover, Dyreng et al. (2010) study CEO characteristics (such as CEO education level, age, tenure, whether he/
she was CFO at a prior rm) as potential determinants of tax
aggressiveness. Their results indicate that none of the CEO
characteristics have a signicant effect on the effective tax rate.
So, more nancial skills due to experience or education do not
seem to increase tax aggressiveness.
Finally, our study intends to broaden our thinking by
incorporating the moderating effect of corporate governance.
Results show that the presence of an outside director in the board
improves the monitoring effectiveness thereby limiting possible
rent extraction behaviour by the CEO. Therefore, these boards
appear to reduce the tax aggressive behaviour of private family
rms where the CEO owns a low amount of shares. Hereby, we
reinforce the ndings of Stockmans, Lybaert, and Voordeckers
(2013) indicating that, in private family rms, governance controls
seem to mitigate the presence of signicant agency conicts.
Previous research has mainly stressed the boards advice role in
family rms. The appointment of independent board members
depends on their ability to provide advice to the family rm based
on their unique knowledge. However, our ndings are in line with
Van den Heuvel, Van Gils, and Voordeckers (2006) and Bammens,
Voordeckers, and Van Gils (2008), indicating that once these
outside board members are appointed, initially for their advice and
knowledge, they do full their legal control task. With respect to
limiting board size or avoiding CEO duality, our results indicate
that these are not effective in mitigating tax aggressive behaviour
by the CEO.
Implications
Several implications can arise from our ndings. Our results
contribute to a better understanding of the determinants of tax
aggressive behaviour in private family rms. Therefore, our
ndings can be of value to shareholders, stakeholders, academics
but also to society as a whole. First, results suggest shareholders
monitoring role to avoid a high extent of tax aggressiveness is
especially important if the CEO has a low amount of shares, which
often occurs in later generation family rms. (Passive) family rm

355

shareholders aiming at keeping up the rms reputation and SEW

may focus on installing an active board of directors consisting of
outside board members that perform their monitoring role
adequately.
Secondly, corporate taxes are also of public concern because tax
aggressive behaviour and the resulting nonpayment of taxes also
have society-wide implications. Taxes play an important role in
funding the provision of public goods, which makes our ndings of
value to tax policymakers who seek to identify the circumstances
that give rise to an increased risk of tax aggressiveness (Lanis &
Richardson, 2013). The ndings of our study indicate that efcient
corporate governance mechanisms, more specically the hiring of
outside independent board members mitigates potential tax
aggressive behaviour and therefore might again give rise to the
question whether law should oblige independent board members in
the context of private (family) rms. Family rms may consider
independent directors as an unnecessary interference in their
decision-making processes and as a potential threat to their power
(Leung, Richardson, & Jaggi, 2014, p. 17). Even though Leung et al.
(2014) nd that independent directors may not contribute to rm
performance, our study does point at the advantages of independent
board members with respect to tax aggressive behaviour of the CEO.
Finally, academics are encouraged to include the socioemotional wealth perspective when conducting research in a family
rm context. Berrone et al. (2012, p. 1) indicate that SEW is the
most important differentiator of the family rm as a unique entity
and, as such, helps explain why family rms behave distinctively.
By complementing the agency view with the SEW view, this study
reveals interesting results. Future research could benet from also
complementing existing theoretical frames with the SEW perspective.
Limitations
Our research also has some limitations that provide many
interesting avenues for future research. Overall, the R-squared values
are relatively low; therefore the ndings need to be interpreted with
care. First, in this study, we only took into account board composition
as a moderator to argue that effective boards can mitigate the
negative relationship between CEO ownership share and tax
aggressiveness. However, other potentially important moderating
effects would be interesting to study. Instead of using board
composition to measure board effectiveness, board processes may be
more suitable. Board processes measure to what extent the board
functions as an effective and controlling board by taking to account
the extent of cognitive conict, effort norms and use of directors
knowledge and skills (Forbes & Milliken, 1999; Zona & Zattoni, 2007).
Moreover, hiring external board members or avoiding CEO
duality, moderators used in our study, can be seen as elements of
board professionalization. While we only focused on these
elements, it would be interesting to investigate whether other
dimensions of rm professionalization (use of more formal control
and HRM systems, professionalization of management and board,
top level activeness, decentralization of authority) can also
contribute to mitigating the negative relation between the CEO
ownership share and tax aggressiveness (for a review, see Dekker
et al., 2012). Finally, besides the board, outside supervision aspects
would also be worth investigating: outside supervision relies on
external auditors and government regulatory bodies to improve
corporate governance (Jin & Lei, 2011).
Secondly, while we take an agency perspective in this article,
the use of other theoretical lenses, such as stewardship theory or
upper echelon theory may complement our study and enrich our
knowledge on tax aggressiveness. In line with Dyreng et al. (2010),
it would be especially interesting to take an upper echelon view
and study the effect of several CEO characteristics (e.g. CEO

356

T. Steijvers, M. Niskanen / Journal of Family Business Strategy 5 (2014) 347357

reputation, experience within the rm and within the industry) on

tax aggressiveness.
Third, in line with other studies on this topic, we use the
effective tax rate as a measure for tax aggressiveness. However, as
tax aggressiveness can be legal or illegal or lie in between (Frank
et al., 2009), it would be interesting if additional measures could be
developed in a private rm context to make the distinction
between legal and illegal tax aggressiveness.
Finally, we coped with a lack of data on certain other rm
characteristics. This lack of data has kept us from extending the focus
on one ownership characteristic, the amount of CEO shares, to the
exact ownership distribution among shareholders. The distinction
between equally distributed ownership and unequal distribution of
ownership (several minority shareholders coupled with one
majority shareholder) or ownership dispersion among non-active
family members versus ownership dispersion among active family
shareholders may affect the shareholder-manager agency problem
and the resulting tax aggressive behaviour. Also with regard to SEW,
we were not able to measure it directly. Therefore, future research
could go one step further by sending out surveys to CEOs in order to
verify the extent to which these noneconomic elements are drivers
of tax aggressive behaviour. The future in this domain lies in the
combination of both economic and noneconomic factors in order to
gain a deeper understanding of tax aggressiveness in private family
rms.
References
Ai, C., & Norton, E. C. (2003). Interaction terms in logit and probit models. Economics
Letters, 80(1), 123129.
Allen, M., & Panian, S. (1982). Power, performance, and succession in the large
corporation. Administrative Science Quarterly, 27(4), 538547.
Anderson, R., Mansi, S., & Reeb, D. (2003). Founding family ownership and the agency
cost of debt. Journal of Financial Economics, 68(2), 263285.
Anderson, R., & Reeb, D. (2003). Founding-family ownership and rm performance:
Evidence from the S&P 500. Journal of Finance, 58(3), 13011328.
Ang, J. S., Cole, R. A., & Lin, J. W. (2000). Agency costs and ownership structure. Journal of
Finance, 55(1), 81106.
Antonakis, J., Bendahan, S., Jacquart, P., & Lalive, R. (2010). On making causal claims: A
review and recommendations. The Leadership Quarterly, 21, 10861120.
Atwood, T., Drake, M., Myers, J., & Myers, L. (2012). Home country tax system
characteristics and corporate tax avoidance: International evidence. The Accounting Review, 87(6), 18311861.
Bammens, Y., Voordeckers, W., & Van Gils, A. (2011). Board of directors in family
businesses: A literature review and research agenda. International Journal of
Management Reviews, 13(2), 134152.
Bammens, Y., Voordeckers, W., & Van Gils, A. (2008). Boards of directors in family rms:
A generational perspective. Small Business Economics, 31(2), 163180.
Banghoj, J., Gabrielsen, G., Petersen, C., & Plenborg, T. (2010). Determinants of executive
compensation in privately held rms. Accounting and Finance, 50(3), 481510.
Beatty, A., & Harris, D. (1999). The effects of taxes, agency costs and information
asymmetry on earnings management: A comparison of public and private rms.
Review of Accounting Studies, 4, 299326.
Bennedsen, M., Nielsen, K., Perez-Gonzalez, F., & Wolfenzon, D. (2007). Inside the
family rm: The role of families in succession decisions and performance. Quarterly
Journal of Economics, 122(2), 647691.
Berenbeim, R. E. (1990). How business families manage the transition from owner to
professional management. Family Business Review, 3(1), 69110.
Berrone, P., Cruz, C., & Gomez-Mejia, L. (2012). Socioemotional wealth in family rms:
Theoretical dimensions, assessment approaches, and agenda for future research.
Family Business Review, 25(3), 258279.
Berrone, P., Cruz, C., Gomez-Mejia, L., & Larraza-Kintana, M. (2010). Socioemotional
wealth and corporate responses to institutional pressures: Do family-controlled
rms pollute less? Administrative Science Quarterly, 55, 82113.
Berry, W. D., DeMeritt, J. H. R., & Esarey, J. (2010). Testing for interaction in binary logit
and probit models: Is a product term essential. American Journal of Political Science,
54(1), 248266.
Blanco-Mazagatos, B., de Quevedo-Puente, E., & Castrillo, L. (2007). The trade-off
between nancial resources and agency costs in the family business: An exploratory study. Family Business Review, 20(3), 199213.
Bliss, M. A. (2011). Does CEO duality constrain board independence? Some evidence
from audit pricing. Accounting and Finance, 51, 361380.
Block, J. (2010). Family management, family ownership, and downsizing: Evidence
from S&P 500 rms. Family Business Review, 23(2), 109130.
Block, J. (2011). How to pay nonfamily managers in large family rms: A principalagent model. Family Business Review, 24, 927.
Bloom, N., & Van Reenen, J. (2007). Measuring and explaining management practices
across rms and countries. Quarterly Journal of Economics, 122(4), 13511408.

Blumentritt, T., Keyt, A., & Astrachan, J. (2007). Creating an environment for successful
nonfamily CEOs: An exploratory study of good principals. Family Business Review,
20(4), 321335.
Brambor, T., Clark, W. R., & Golder, M. (2006). Understanding interaction models:
Improving empirical analyses. Political Analysis, 14(1), 6382.
Burgstahler, D., Hail, L., & Leuz, C. (2006). The importance of reporting incentives:
Earnings management in European private and public rms. The Accounting
Review, 81(5), 9831016.
Carcello, J. V., Hermanson, D. R., Neal, T. L., & Riley, R. A., Jr. (2002). Board characteristics
and audit fees. Contemporary Accounting Research, 196, 365384.
Chen, S., Chen, X., Cheng, Q., & Shevlin, T. (2010). Are family rms more tax aggressive
than non-family rms? Journal of Financial Economics, 95, 4161.
Chrisman, J., Chua, J., Pearson, A., & Barnett, T. (2010). Family involvement, family
inuence and family-centered non economic goals in small rms. Entrepreneurship
Theory and Practice, 127.
Chrisman, J., Chua, J., & Sharma, P. (2005). Trends and directions in the development of
a strategic management theory of the family rm. Entrepreneurship Theory and
Practice, 29, 555575.
Chua, J., Chrisman, J., & Sharma, P. (1999). Dening the family business by behavior.
Entrepreneurship Theory and Practice, 23, 1939.
Chua, J., Chrisman, J., & Sharma, P. (2003). Succession and nonsuccession concerns of
family rms and agency relationship with nonfamily managers. Family Business
Review, 16(2), 89107.
Chyz, J., Leung, W., Li, O., & Rui, O. (2013). Labor unions and tax aggressiveness. Journal
of Financial Economics, 108, 675698.
Classen, N., Van Gils, A., Bammens, Y., & Carree, M. (2012). Accessing resources from
innovation partners: The search breadth of family SMEs. Journal of Small Business
Management, 50(2), 191215.
Corbetta, G. (1995). Patterns of development of family businesses in Italy. Family
Business Review, 8(4), 255265.
Dalton, D., Daily, C., Johnson, J., & Ellstrand, A. (1999). Number of directors and nancial
performance: A meta-analysis. Academy of Management Journal, 42, 674686.
Dekker, J. C., Lybaert, N., Steijvers, T., Depaire, B., & Mercken, R. (2012). Family rm
types based on the professionalization construct: Exploratory research. Family
Business Review, 26(1), 8199.
Desai, M., & Dharmapala, D. (2006). Corporate tax avoidance and high-powered
incentives. Journal of Financial Economics, 84, 591623.
Dyer, W. G. (1988). Culture and continuity in family rms. Family Business Review, 1(1),
3750.
Dyer, W. G., & Whetten, D. (2006). Family rms and social responsibility:
Preliminary evidence from the S&P500. Entrepreneurship Theory and Practice,
785802.
Dyreng, S., Hanlon, M., & Maydew, E. (2010). The effects of executives on corporate tax
avoidance. The Accounting Review, 85(4), 11631189.
Fama, E. F., & Jensen, M. C. (1983). Separation of ownership and control. Journal of Law
and Economics, 26, 301325.
Finer, L., Laine, M., & Ylonen, M. (2012). Verosuunnittelu vauhdittaa Enson sellusampoa. Talouselama, 23, 3843.
Forbes, D., & Milliken, F. (1999). Cognition and corporate governance: Understanding
boards of directors as strategic decision making groups. Academy of Management
Review, 24, 489505.
Frank, M., Lynch, L., & Rego, S. (2009). Tax reporting aggressiveness and its relation to
aggressive nancial reporting. The Accounting Review, 84, 467496.
Gedajlovic, E., & Carney, M. (2010). Market, hierarchies and families: Toward a
transaction cost theory of the family rm. Entrepreneurship Theory and Practice,
34, 11451171.
Gersick, K., Lansberg, I., Desjardins, M., & Dunn, B. (1999). Stages and transitions:
Managing change in the family business. Family Business Review, 12(4), 287297.
Gersick, K., Davis, J., Hampton, M., & Lansberg, I. (1997). Generation to generation: Life
cycles of the family business. Boston, MA: Harvard Business School Press.
Gomez-Mejia, L. R., Cruz, C., Berrone, P., & De Castro, J. O. (2011). The bind that ties:
Socioemotional wealth preservation in family rms. The Academy of Management
Annals, 5(1), 653707.
Gomez-Mejia, L., Haynes, K., Nunez-Nickel, M., Jacobson, K., & Moyano-Fuentes, J.
(2007). Socioemotional wealth and business risks in family-controlled
rms: Evidence from Spanish olive mills. Administrative Science Quarterly, 52,
106137.
Gujarati, D. N. (1995). Basic econometrics (3rd ed.). New York: McGraw-Hill.
Hambrick, D. C., & Mason, P. A. (1984). Upper echelons: The organization as a reection
of its top managers. Academy of Management Review, 9, 93106.
Hanlon, M., & Heitzman, S. (2010). A review of tax research. Journal of Accounting and
Economics, 50, 127178.
Hanlon, M., & Shevlin, T. (2005). Book-tax conformity for corporate income: An
introduction to the issues. Tax Policy and the Economy, 19, 101134.
Harford, J., Mansi, S., & Maxwell, W. (2008). Corporate governance and rm cash
holdings in the US. Journal of Financial Economics, 87, 535555.
Huseynov, F., & Klamm, B. (2012). Tax avoidance, tax management and corporate social
responsibility. Journal of Corporate Finance, 18, 804827.
Jensen, M. C. (1993). The modern industrial revolution, exit, and the failure of internal
control systems. Journal of Finance, 48, 831880.
Jensen, M. (1986). Agency costs of free cash ow, corporate nance and takeovers.
American Economic Review, 76(2), 323329.
Jensen, M., & Meckling, W. (1976). Theory of the rm: Managerial behavior, agency
costs and ownership structure. Journal of Financial Economics, 3, 305360.
Jin, X., & Lei, G. (2011). Audit supervision, property of ultimate controller and tax
aggressiveness. Auditing Research, 5, 8.

T. Steijvers, M. Niskanen / Journal of Family Business Strategy 5 (2014) 347357

Johannisson, B., & Huse, M. (2000). Recruiting outside board members in the small
family business: An ideological challenge. Entrepreneurship & Regional Development, 12, 353378.
Kam, C., & Franzese, R. (2007). Modeling and interpreting interactive hypotheses in
regression analysis. Michigan: University of Michigan Press.
Karra, N., Tracey, P., & Phillips, N. (2006, November). Altruism and agency in the family
rm: Exploring the role of family, kinship, and ethnicity. Entrepreneurship Theory &
Practice, 861877.
Lanis, R., & Richardson, G. (2011). The effect of board of director composition on
corporate tax aggressiveness. Journal of Accounting and Public Policy, 30, 5070.
Lanis, R., & Richardson, G. (2013). Corporate social responsibility and tax aggressiveness: A test of legitimacy theory. Accounting, Auditing & Accountability Journal,
(1), 75100.
Leung, S., Richardson, G., & Jaggi, B. (2014). Corporate board and board committee
independence, rm performance, and family ownership concentration: An analysis
based on Hong Kong rms. Journal of Contemporary Accounting & Economics, 10,
1631.
Lubatkin, M., Schulze, W., Ling, Y., & Dino, R. (2005). The effects of parental altruism on
the governance of family-managed rms. Journal of Organizational Behavior, 26,
313330.
Maciejovsky, B., Schwarzenberger, H., & Kirchler, E. (2012). Rationality versus
emotions: The case of tax ethics and compliance. Journal of Business Ethics,
109(3), 339350.
Manseld, E. R., & Helms, B. P. (1982). Detecting multicollinearity. American Statistician,
36(3), 158.
Michielsen, A., Voordeckers, W., Lybaert, N., & Steijvers, T. (2013). CEO compensation in
private family rms: Pay-for-performance sensitivity and the moderating role of
ownership and management conditions. Family Business Review, 26(2), 140160.
Miller, D., & Le Breton-Miller, I. (2006). Family governance and rm performance:
Agency, stewardship, and capabilities. Family Business Review, 19(1), 7387.
Mills, L., & Newberry, K. (2001). The inuence of tax and non-tax costs on book-tax
reporting differences: Public and private rms. Journal of American Tax Association,
23, 119.
Minnick, K., & Noga, T. (2010). Do corporate governance characteristics inuence tax
management? Journal of Corporate Finance, 16, 703718.
Molero, J. C., & Pujol, F. (2012). Walking inside the potential tax evaders mind: Tax
morale does matter. Journal of Business Ethics, 105, 151162.
Niskanen, J., & Keloharju, M. (2000). Earning cosmetics in a tax-driven accounting
environment: Evidence from Finnish public rms. European Accounting Review,
9(3), 443452.
Norton, E. C., Wang, H., & Ai, C. (2004). Computing interaction effects and standard
errors in logit and probit models. Stata Journal, 4(2), 13.
Salvato, C., & Moores, K. (2010). Research in accounting in family rms: Past accomplishments and future challenges. Family Business Review, 23(3), 193215.
Scholes, M., Wolfson, M., Erickson, M., Maydew, E., & Shevlin, T. (2005). Taxes and
business strategy: A planning approach. New York/Upper Saddle River: Pearson/
Prentice Hall.

357

Schulze, W. S., Lubatkin, M., & Dino, R. (2003a). Toward a theory of agency and altruism
in family rms. Journal of Business Venturing, 18(4), 473490.
Schulze, W. S., Lubatkin, M., & Dino, R. (2003b). Exploring the agency consequences of
ownership dispersion among the directors of private family rms. Academy of
Management Journal, 46(2), 179194.
Shackelford, D., & Shevlin, T. (2001). Empirical tax research in accounting. Journal of
Accounting and Economics, 31, 321387.
Sharma, P., & Manikuti, S. (2005). Strategic divestments in family rms: Role of
family structure and community culture. Entrepreneurship Theory and Practice,
29, 293311.
Songini, L., Gnan, L., & Malmi, T. (2013). The role and impact of accounting in family
business. Journal of Family Business Strategy, 4, 7183.
Steijvers, T., & Niskanen, M. (2013). The determinants of cash holdings in private family
rms. Accounting and Finance, 53(2), 537560.
Stockmans, A., Lybaert, N., & Voordeckers, W. (2013). The conditional nature of board
characteristics in constraining earnings management in private family rms.
Journal of Family Business Strategy, 4, 8492.
Stockmans, A., Lybaert, N., & Voordeckers, W. (2010). Socioemotional wealth and
earnings management in private family rms. Family Business Review, 23(3),
280294.
Tsai, W., Hung, J., Kuo, Y., & Kuo, L. (2006). CEO tenure in Taiwanese family
and nonfamily rms: An agency theory perspective. Family Business Review,
19(1), 1128.
Van den Berghe, L. A., & Carchon, S. (2003). Agency relations within the family business
system: An exploratory approach. Corporate Governance: An International Review,
11(3), 171179.
Van den Heuvel, J., Van Gils, A., & Voordeckers, W. (2006). Board roles in small and
medium-sized family businesses: Performance and importance. Corporate Governance: An International Review, 14(5), 467485.
Van Tendeloo, B., & Vanstraelen, A. (2008). Earnings management and audit quality in
Europe: Evidence from the private client segment market. European Accounting
Review, 17(3), 447469.
Westhead, P. (2003). Company performance and objectives reported by rst and multigeneration family companies: A research note. Journal of Small Business and
Enterprise Development, 10, 93105.
Westhead, P., & Howorth, C. (2007). Types of private family rms: An exploratory
conceptual and empirical analysis. Entrepreneurship & Regional Development, 19,
405431.
Yuan, G., McIver, R., & Burrow, M. (2012). Corporate income tax aggressiveness in
China: Regulatory environment and ownership impact. Journal of Business Management, 5, 144160.
Zona, F., & Zattoni, A. (2007). Beyond the black box of demography: Board processes
and task effectiveness within Italian rms. Corporate Governance: An International
Review, 15(5), 852864.
Zona, F., Minoja, M., & Coda, V. (2013). Antecedents of corporate scandals: CEOs
personal traits, stakeholders cohesion, managerial fraud, and imbalanced corporate strategy. Journal of Business Ethics, 113, 265283.